How we rank, and where our data comes from

You may not agree with this order. We present the data as we find it, and we say plainly how much we think each measure is worth. Every weight is published, every figure is sourced and dated, and where we cannot prove something we leave it blank rather than fill it in. Complete transparency is the whole claim — disagree with the weighting, and you can see exactly what you are disagreeing with.

39 subjects and 27 universities, built from official government data and nothing else. This is every measure, every weight, and where each figure came from — including the ones we do not have.

What makes this different

Standing is measured, not surveyed. Academic and professional standing count here — but through REF 2021, where UK academics formally peer-review each other's research, and through whether a professional body licenses the degree as a route into the job. Nobody is asked for an opinion.
Entry grades count for a little, and only for one reason. A selective course assembles a stronger cohort, and who you spend three years arguing with is part of what you get. That is worth something — but it measures the intake, never what the course does with it, so it sits below every outcome measure here and at a third of the weight the Complete University Guide gives it.
Tiers, not places. Every other table prints a number and lets you assume the gap between 17th and 18th means something. We group the strongest half-dozen for a subject together and list them alphabetically, because the order inside that group is not something the data can support.
A missing figure is never counted as a bad one. Where a university has no data for a measure it is judged on what exists, the gap reads “not published” rather than as a low number, and a course with too little data to judge fairly is listed separately instead of being ranked badly.

What goes into the score

Each is measured against its own subject first, because £31,500 is strong for History and ordinary for Medicine. Research quality carries a confidence per subject rather than one blanket caveat: Economics is assessed by a research unit of exactly that name, while Psychology and Biological Sciences share theirs with neighbouring fields and are marked as the weaker reading they are. Professional accreditation is only scored in the 7 subjects where the profession genuinely routes through the degree — in Medicine every course is approved and in Classics none is, so there it separates nobody.

Measure	Weight
Student satisfaction (all seven official NSS themes)	15%
Research quality — REF 2021, UK academics peer-reviewing research	18%
Recognised by a professional body as a route into the profession	5%
Average UCAS points of the students who got in	20%
Earnings five years on, against the national figure for the subject	13%
In work or further study 15 months after graduating	5%
Share in professional or managerial roles	10%
Students who continue past first year	14%

Weights hashed as c6abef4e77df into the data snapshot, so a change to them is as auditable as a change to the data.

What we leave out, and why

Opinion polls of academics and employers: The QS World University Rankings take 30% of their score from an academic reputation survey and 15% from an employer one — 45% of the result is what people think, and what people think lags reality by a generation. We count the same two things a different way: REF 2021 is UK academics assessing research with a published method and named panels, and professional accreditation is a body like the IET or the SRA putting its name to a degree as a route into the profession. Together they carry 15% against QS's 45%. Reputation you can check beats reputation you are asked to trust.
Entry standards as a main ingredient: In the Complete University Guide, entry standards carry the joint-highest weight, so a course becomes "better" by rejecting more people. We count them, because a stronger intake really does change what three years there is like — but at a fraction of that, and below every measure of what actually happened to graduates. A table led by selectivity tells you who applied, not what they got.
Going on to a master's: We hold the figure and deliberately do not score it. Postgraduate study is not better than a job, so rewarding it would be us asserting a value judgement the data does not support. It stays available as a filter, never as a mark.
Spending per student, staff-to-student ratios: Measures of what a university spends, not of what its students get. We would rather measure the outcome directly, and we can.

What the order can and cannot tell you

Nearly half of National Student Survey results are withheld for small sample sizes, and where a figure does survive it can rest on a few dozen responses. A twenty-person cohort having a good year can top a table on chance alone, so we pull every small-sample figure toward its subject average in proportion to how little data sits behind it — the reason these numbers move only when a new official release lands, never week to week.

What survives that is enough to say which group a university belongs in, and not enough to order the group. When we changed one defensible weighting choice, courses moved an average of four points — wider than most gaps between neighbours. So we publish tiers, each one ending at the widest real separation available rather than after a fixed count, and inside a tier the names are alphabetical. Open any row and it names who it is too close to call against. Nobody else shows you that, and it is the difference between a ranking you can argue with and one you have to take on trust.

An honest word about who we cover

The tiers are cut by comparing each university against all UK courses in that subject, not just the 27 we list. Those 27 are not a random sample of British higher education, so most of them land in the top half nationally — of 761 university-subject entries, 440 sit in the top two fifths of the country and 10 in the bottom fifth.

Which is why a tier is a group and not a placing. A university in the third tier here can still be ahead of most courses in the country in its subject, so printing a number against the others on this page would tell you the less useful of the two facts. Open a university instead and you get the published figures themselves.

Last updated: 1 August 2026

In one minute

Your child ranks what matters to them — earnings, teaching quality, how selective a place is, and five other things. That ranking is the only thing that decides the order courses appear in.
We supply the facts, not the priorities. Every figure on a course card is published by the government, UCAS, or the university itself. We have not scored anything ourselves.
Where we do not know, we leave it blank. Not rarely — student satisfaction is missing on 272 of our 5,412 courses. We show you which ones and why, further down. That is the whole difference between this and a league table.
No university has paid to be here, and none can move itself up.

Every number Writ shows you about a course comes from an official public dataset. Not from a magazine, not from a survey we ran, and not from a university's marketing. This page lists all of them, the years they cover, and exactly what we do with them.

We publish this because we are asking you to trust an ordering of courses. A league table asks for that trust and shows you a single score it will not explain. We would rather tell you where the numbers come from, what they can tell you, and where the honest answer is that we do not know.

The sources

Discover Uni (the Unistats dataset)

The government's official comparison site for UK degrees. Every university has to submit to it by law.

Published by: Office for Students / HESA, on behalf of the UK funding bodies
Covers: 30,562 courses at 495 UK providers
We retrieved it: 29 September 2026

The official course-level dataset every UK university is required to submit to. It carries graduate earnings, employment, continuation and student satisfaction for individual courses where enough students were counted, and for the subject at that university where they were not, saying which.

How we use it: This is the spine. Four of the six figures a course card can show come from here.

Longitudinal Education Outcomes (LEO)

What graduates in a subject at each university actually earned, taken from their tax records rather than from a survey.

Published by: Department for Education, from HMRC tax and DWP benefits records
Covers: Earnings three and five years after graduation
We retrieved it: 29 September 2026, via the Discover Uni dataset

Actual earnings taken from tax records — not a graduate survey, and not what universities say their graduates earn. It is the hardest evidence on earnings that exists in UK higher education.

How we use it: The 'Earnings' line on a card. It is published for each subject at each university, never for one course, so every course in that subject there shares it. We show the difference between it and the national median for the same subject, so you are comparing like with like.

National Student Survey (NSS)

The survey every final-year student in the UK is asked to fill in about their own course.

Published by: Office for Students
Covers: Final-year undergraduates, all seven official themes
We retrieved it: 29 September 2026, via the Discover Uni dataset

What students on the course actually said about teaching, assessment and feedback, academic support, organisation, learning resources, learning opportunities and student voice.

How we use it: The 'Satisfaction' line. We use all seven official themes rather than picking a few flattering questions.

Graduate Outcomes

A national survey asking graduates what they are actually doing, fifteen months after they finished.

Published by: HESA
Covers: Graduates surveyed 15 months after finishing
We retrieved it: 29 September 2026, via the Discover Uni dataset

Whether graduates are in work or further study, and what proportion are in professional or managerial roles.

How we use it: The 'Career outcomes' and 'In work or study' lines.

UCAS end-of-cycle data

UCAS's own record of how many people applied to each university in each subject, and how many were offered a place.

Published by: UCAS
Covers: 2025 admissions cycle (the file also holds 2019 onward)
We retrieved it: 31 July 2026

How many applications each university received in each subject area, and how many of them received an offer.

How we use it: The 'Offer rate' line and the competitiveness badge. Read the caveat below — this is per subject area, not per course.

Research Excellence Framework 2021 (REF)

A panel of academics reading and grading the research a university department actually produced.

Published by: Research England and the other UK funding bodies
Covers: The 2021 assessment — the most recent one; the next is expected around 2029
We retrieved it: 30 July 2026

Expert peer review of every university's research output, graded from unclassified to four star.

How we use it: The 'Research' line, as a score out of 4. Always marked approximate — see below.

The universities' own course pages

The course page on the university's own website — the same page you would find yourself.

Published by: Each university
Covers: 3,514 courses linked to the university's own page (3,200 with module lists taken from it), across 27 universities
We retrieved it: Continuously; each course records its own fetch date

Course titles, entry requirements, module lists and summaries, taken from the university's own published page for that course.

How we use it: Everything descriptive on a card. Every course links out to the university's page, never to an aggregator.

How we turn that into an order

In onboarding you rank eight things by how much they matter to you. That ranking, and nothing else, decides the order courses appear in. There is no Writ score, no editorial weighting, and no way for a university to pay to move up.

All eight are backed by real data. Location is the one that needs something from you first: if it lands in your top three, we ask which parts of the country you would actually consider, and courses there rise. We would rather ask than guess where someone wants to live.

Every figure is compared within its own subject. A £30,000 median salary is strong for History and modest for Medicine, so comparing raw salaries across subjects would just sort your list by subject rather than by quality.

Where the blanks are

Six figures can appear on a course card. None of them is present for every course, because the official statistics suppress a figure when the group behind it is small enough that a student could be identified from it. This is exactly how often each one is there.

How often we have each figure

5,412 courses

Earnings, against the subject median5,094 of 5,41294%
5,094 subject-wide · 318 left blank
Research quality (REF 2021)4,791 of 5,41289%
4,791 subject-wide · 621 left blank
Students who continue past first year5,141 of 5,41295%
2,764 subject-wide · 271 left blank
Student satisfaction (NSS)5,140 of 5,41295%
3,123 subject-wide · 272 left blank
In work or study, 15 months on4,854 of 5,41290%
3,708 subject-wide · 558 left blank
In professional or managerial roles4,809 of 5,41289%
3,871 subject-wide · 603 left blank

Where too few students on a course were counted, Discover Uni publishes the figure for that subject at the university instead, and says so. We do the same: the fainter part of each bar is those subject-wide figures, and every one is marked “subject-wide” on the course. Earnings and research are only ever published per subject, so those two bars are subject-wide throughout. A figure from a group of fewer than 20 students is shown with its size and no verdict, and the rankings leave it out. A blank means nothing was published at all. A league table would fill it with a national average. We show you the space.

Some of it is ours, though, and leaving that out would be the kind of thing this page exists not to do. 260 of our 5,412 courses are not yet matched to their official record, and each of those is a blank on every bar above. Most are joint degrees the official dataset files under one half, or degrees it does not publish separately at all. Closing every one of them would move the top bar to about 100% and the bottom to about 93% — not to 100%, because the rest was never published.

What we will not do

We never guess.
Our course list and the official datasets name courses differently. Where we cannot prove two entries are the same degree, that course shows no figures at all rather than borrowing a similar course's numbers. Putting one degree's graduate salary on another is the worst thing a tool like this could do, so a blank space is the honest answer and you will see a few of them.
We never fill a gap with an average.
Where a figure was withheld because too few students were involved, the space stays empty. We do not substitute the subject average, the university average, or last year's number.
No university pays to appear here.
There is no sponsored placement anywhere in the course finder, and no university can influence its position.
Your priorities never reach the writing coach.
What matters to you when choosing a course is not what should shape a personal statement. If the coach knew you ranked earnings first it would start steering your statement toward justifying a salary, which is exactly what admissions tutors mark down. That separation is enforced in the software, not just intended.

What these numbers can and cannot tell you

Offer rates describe a subject area, not one course.
UCAS publishes how many offers each university makes in each broad subject area, not for individual degrees. So every History course at a university shares a figure with every Philosophy course there. It is a real measure of how selective that university is in that area. It is not a prediction of your chances on one specific course.
Earnings describe people who graduated years ago.
Earnings are measured three and five years after graduation, so the most recent figures describe students who started university around a decade ago. That is what it means to measure real outcomes rather than promises — but it is history, not a forecast.
Research quality is matched carefully but not perfectly.
Research is assessed in its own categories, which do not line up exactly with the subject groupings used everywhere else. We map them as closely as we can and mark every research figure as approximate, because it is a weaker claim than the others.
Entry requirements are typical, not binding.
Published offers vary by college, by contextual circumstances, by admissions test performance and by year. Always confirm on the university's own page, which every course card links to.

Found something wrong?

Open any course and use the report link inside it, or email enquiries@writai.co.uk. Corrections to course data are checked against the university's own page before anything changes.

Writ is not affiliated with, endorsed by, or connected to UCAS, HESA, the Office for Students, the Department for Education, Research England or Discover Uni. We use their published open data under their terms. See our Terms and Privacy Policy.